Agreement is found over the entire range for a reasonably small area of the sea References and Notes

of wind speeds (14 to 34 m/sec) derived
for the rain-free portions of this pass.
There is no evidence of any saturation of
the radar cross section with increasing
wind speed.
Conclusion
The accuracy of the near-surface wind
speeds derived from the P-3 measure-
ments is ± 10 percent, with errors due
in part to the inertial navigation system
measurement itself and to approxima-
tions inherent in the hurricane boundary
layer model. There may also be small
errors in the measurement and process-
ing of the remotely sensed data. Consid-
ering these errors and possible misregis-
tration of the two aircraft tracks, certain
discrepancies are to be expected. How-
ever, when the remotely sensed wind
speed, wind direction, and rain rate
trends are compared with the values
from in situ sources, there is close agree-
ment.
In the future, more sophisticated
beam-aiming techniques will be used to
form a more optimum sampling scheme.
Judicious choice of altitude and inci-
dence angle and adjustment of the anten-
na azimuth sweep to compensate for the
aircraft's speed will allow measurement
of a° over the desired range of azimuths
surface. This would be of considerable
value in a hurricane, where numerous
meteorological phenomena are distribut-
ed over relatively short distances. An-
other improvement would be accurate
detection of, and eventual correction for,
atmospheric attenuation within the radar
beam path.
The data presented here are based on
the first of a total of approximately 20
multi-aircraft passes through Hurricane
Allen's eye on 5 and 8 August. On many
of these passes the AMSCAT antenna
was programmed to obtain azimuth
sweeps at various incidence angles other
than 400, and also to obtain elevation
sweeps at fixed azimuths. The SFMR
was operated in a multifrequency mode
on all but one pass. In situ measurements
are available for nearly all passes.
The data from most of these passes
will be analyzed with a view toward
optimizing the sensor operating modes
and the algorithms for future hurricane
experiments. Also, an attempt will be
made to process on-board (in real time)
at least the SFMR data to surface wind
speed and rain rate.
The results of this experiment demon-
strate the feasibility of microwave re-
mote sensing techniques to obtain from
high altitudes information which is pres-
ently obtained at low altitudes and at
considerable risk.
1. R. C. Gentry, J. Underwater Sci. Technol. 2,
204 (1970).
2. G. D. Atkinson and C. R. Holliday, Mon.
Weather Rev. 105, 421 (1977).
3. D. E. Barrick and C. T. Swift, IEEE J. Oceanic
Eng. 5 (No. 2), 74 (1980).
4. D. B. Ross, J. Conaway, V. J. Cardone, IEEE
Trans. Geosci. Electron. 8 (No. 4), 326 (1970).
5. W. J. Nordberg, J. Conaway, D. B. Ross, T.
Wilheit, J. Atmos. Sci. 28, 429 (1971).
6. D. B. Ross and V. J. Cardone, J. Geophys. Res.
79, 444 (1974).
7. N. W. Guinard, J. T. Ransone, Jr., J. C. Daley,
ibid. 76, 1525 (1971).
8. K. Krishen, ibid., p. 6528.
9. W. L. Jones, L. C. Schroeder, J. L. Mitchell,
IEEE J. Oceanic Eng. 2 (No. 1), 52 (1977).
10. D. B. Ross, B. Chu, W. Brown, J. McFadden,
paper presented at the Remote Sensing Sympo-
sium, Willow Run Laboratory, Ann Arbor,
Mich., 15-20 April 1974.
11. D. E. Weissman, D. King, T. W. Thompson, J.
Appl. Meteorol. 18, 1023 (1979).
12. W. J. Cardone et al., NASA Contract. Rep. CR-
147487 (1976).
13. D. P. Jorgenson and P. T. Willis, unpublished
manuscript.
14. M. S. Moss and F. J. Merceret, Mon. Weather
Rev. 104, 967 (1975).
15. M. D. Powell, ibid. 108, 757 (1980).
16. R. F. Harrington, NASA Tech. Memo. 81847
(1980).
17. W. J. Webster, T. T. Wilheit, D. B. Ross, P.
Gloersen, J. Geophys. Res. 81, 3095 (1976).
18. L. J. Battan, Radar Observation of the Atmo-
sphere (Univ. of Chicago Press, Chicago, 1973),
p. 73.
19. P. Gloersen and F. T. Barath, IEEE J. Oceanic
Eng. 2 (No. 2), 172 (1977).
20. W. L. Jones, F. J. Wentz, L. C. Schroeder, J.
Spacecraft Rockets 15 (No. 6), 368 (1978).
21. L. C. Schroeder, D. H. Boggs, G. Dome, 1. M.
Halberstam, W. L. Jones, W. J. Pierson, F. J.
Wentz, unpublished manuscript.
22. V. J. Cardone, "Specification of the wind distri-
bution in the marine boundary layer for wave
forecasting" (Report TR 69-1, School of Engi-
neering and Science, New York University,
University Heights, N.Y., 1970).
23. We thank K. Kiley and F. Marks for helping to
prepare Figs. 3 and 5. We also acknowledge R.
Couch and R. Harington for instrument sup-
port.

drugs, toxins, pheromones, odors, and

Tandem Mass Spectrometry
Fred W. McLafferty
"Needle in a haystack" analytical agricultural products, drugs, and explo-
problems have become important in a sives.
variety of areas. Specific compounds Specificity, accuracy, sensitivity, and
which must be recognized in the pres- speed are the key performance charac-
ence of many others include drugs or teristics that must justify the cost of the
disease-indicative compounds in biologi- analytical technique chosen. Specificity
cal fluids, pollutants in environmental must be sufficient to distinguish the com-
samples, chemicals used in the classifi- pound sought (the targeted molecule)
cation of plant and animal species (che- from all others present. For quantitation,
motaxonomy), natural insect attractants the specific instrument reponse must be
(pheromones), flavor- and odor-produc- directly related to the amount of targeted
ing compounds, petroleum and synthet- compound present, independent of other
ic fuels, process control compounds, components. Sensitivity is a limiting fac-
chemical warfare agents, and contraband tor for many applications; enzymes,
280 0036-8075/81/1016-0280$01.00/0 Copyright © 1981 AAAS
flavors can be active in picogram con-
centrations. Speed and applicability are
important, for a laboratory's willingness
to invest in equipment and trained per-
sonnel for a technique will depend on the
number and variety of analytical prob-
lems it will solve.
Radioimmunoassay (RIA) is a prime
example of a technique that combines
unusual selectivity and sensitivity (1);
these properties have led to its wide-
spread use in clinical and research labo-
ratories around the world. For example,
o- 13 gram of gastrin per milliliter is
readily measurable, and steroid mole-
cules differing only in the presence of a
hydroxyl group can be distinguished by
RIA. Its technical simplicity makes pos-
sible a low unit analytical cost and high
accuracy with nonprofessional person-
nel, although several hours are required
for an analysis and extensive develop-
ment of methods may be necessary for a
new targeted molecule.
The author is a professor of chemistry at Cornell
University, Ithaca, New York 14853.
SSCIENCE, VOL. 214,16 OCTOBER 1981
 In contrast, ion-selective electrodes and collection of statistically meaningful to analyze one or more samples per hour
(2, 3) respond nearly instantaneously to data. Relatively high specificity can be for a multiplicity of components per sam-
changes in the concentration of the tar- achieved by use of the hundreds of possi- ple. Many of the problems arising from
geted molecule, yielding analytical infor- ble mass positions that are completely the technical complexity of the instru-
mation at an unusually low cost. Howev- resolved from their neighbors, combined mentation have been met by using a
er, their specificity and sensitivity are, in with selective ionization of targeted dedicated computer to handle data re-
general, substantially lower than those of compounds. However, a complex mix- duction and instrument control, which
RIA. An electrode specific for dodecyl- ture such as a blood plasma can produce also simplifies application to new prob-
lems. In addition to targeted compound
analysis, GC-MS has been uniquely ap-
Summary. Coupling mass spectrometers in series provides a new technique that plicable to the identification of total un-
has many advantages for the analysis of specific organic compounds in complex knowns in complex mixtures such as
mixtures. Sensitivity to picograms of targeted compounds can be achieved with high pollutants, pheromones, and flavors (8).
specificity and nearly instantaneous response. The targeted compound is selectively The additional specificity provided by a
ionized, and its characteristic ions are separated from most others of the mixture in chromatographic preseparation greatly
the first mass spectrometer. The selected primary ions are then decomposed by reduces interference without appreciably
collision, and from the resulting products the final mass analyzer selects secondary reducing the sensitivity of MS, but the
ions characteristic of the targeted compound. Tandem mass spectrometry can speed of analysis is limited by the time
achieve specificities and sensitivities equivalent to those of methods such as required for the chromatograph.
radioimmunoassay and gas chromatography/mass spectrometry, while performing A preseparation can also be provided
analyses in much shorter times. by use of an additional mass spectrome-
ter, and this technique is known as tan-
dem mass spectrometry, or MS-MS (13,
trimethylammonium ion gives a mini- a peak at virtually every mass position 14). Such tandem mass spectrometers
mum detectable signal for -10-9 g, and from m/z (mass-to-charge ratio) 50 to (Figs. 1 and 2) can have much greater
to avoid interference other tetraalkylam- 500, interfering with the detection of selectivity than a single mass spectrome-
monium ions present must differ in mo- almost any trace component with a peak ter without serious loss of sensitivity,
lecular weight by -70 (3). Analytical in this mass range. and retain the unique advantage of a
methods based on enzymatic reactions One of the most broadly applicable response time < lo-3 second. Except for
can be unusually specific and sensitive techniques for specific trace analysis in scattered early reports (15-18), most re-
(4); when combined with ion-selective complex mixtures is GC-MS, the tandem search activity in this field has occurred
electrodes, they can also provide a fast coupling of gas chromatography and in the last 5 years (19-25). In this article I
response. However, these methods are mass spectrometry (8). This combines discuss basic aspects of the technique
applicable only to relatively specialized the complementary specificities of the and recent developments.
chemical systems, as are a variety of chromatographic and MS separations
other sensitive detectors (5, 6). with the subpicogram sensitivity and
Two-dimensional chromatography is a quantitative accuracy of MS. Only com- Principles
broadly applicable method (5) by which pounds that are sufficiently volatile for
selectivity can be increased. To separate GC can be used in GC-MS, but substitut- The hypothetical complex mixture
a targeted molecule from 10,000 related ing a liquid chromatograph (LC) (9) ex- ABCD (the molecule targeted for analy-
compounds, it is usually easier to find tends the sample volatility range to that sis), BCDA, CBDA, EFGH, IJKL, . . .
chromatographic conditions for separat- of MS. The high instrument and person- is introduced continuously into the MS
ing an individual fraction containing the nel training costs are offset by the ability ionizing region to produce a complex
targeted compound plus -100 others,
and then find other conditions for sepa-
rating the targeted compound from these
100. Analysis for other compounds in the
sample could require additional separa-
tions, increasing the sample analysis
time as well as the time needed to opti-
mize each set of separation conditions.
Mass spectrometry (MS) (7-9) has a
been for many years an analytical tool -

of unusually high sensitivity and speed

coupled with good specificity for com- 'a0
pounds with molecular weights up to
- 1000, and recent research has raised
this limit substantially (10-12). The elec-
tron multiplier makes possible the detec-
tion of a single ion, so that sensitivities of
10-'3 to 10-15 g are attainable with com- Ionizer Analyzer
mercial instruments (8). Response can be
virtually instantaneous, as ion formation Turbomolcular
and transmission require < 10-3 second; pumps
the speed of an analysis is determined by Fig. 1. Schematic of a tandem quadrupole MS-MS instrument. [Courtesy of the Finnigan
the time required for sample introduction Corporation]
16 OCTOBER 1981 281
mixture of ions, among them the isomers
ABCD+, BCDA+, and CBDA+. The
first mass spectrometer (MS-I) is set to
transmit only ions whose m/z value cor-
responds to ABCD+, so that these and
BCDA+ and CBDA+ ions will exit con-
tinuously from MS-I. Energy is added to
these ions by collision, causing them to
decompose; the secondary ions pro-
duced are then separated by MS-1I. If
only ABCD+ primary ions (not BCDA+
or CBDA+) yield the secondary ions
AB+, MS-1I separation and detection of
the latter will provide a direct measure of
the original concentration of ABCD in
the sample mixture. As in multiple ion
monitoring in GC-MS, other peaks in the
secondary mass spectrum of a primary
ion can be used to increase the specific-
ity of measuring its corresponding com-
ponent. For example, if the abundances
of the secondary ions ABC+ and AB'
are in the ratio found when pure ABCD
is introduced into the MS-MS, it is much
less likely that these ions were formed by
components other than ABCD in the
original mixture. Analysis of other com-
ponents of the mixture is straightfor-
ward; measuring EFGH would involve
separating the EFGH+ ions by MS-I and
separating and detecting collision-pro-
duced secondary ions such as EFG+ and
FGH+ by MS-II.
Selective ionization. For mass spec-
trometry, convenient and sensitive ion-
ization methods have been developed
that are selective for a variety of com-
pound types (26). Ionization with 70-eV
electrons (electron ionization or El), the
method used most commonly for com-
143 mm
-1.--407 mm-A
589 mm ESA-1
MS-I
MS-I entrance slit
Ion source
Fig. 2. Tandem double-focusing MS-MS instrument constructed at Cornell.
282
pound identification in GC-MS, is rela-
tively undesirable for MS-MS analysis of
targeted compounds, as it produced doz-
ens of primary ions from most com-
pounds. Chemical ionization (CI) is
probably the most versatile selective ion-
ization technique; a wide range of ioniz-
ing reagent gases and conditions that
produce negative as well as positive ions
have been recommended for many com-
mon classes of organic compounds (26,
27). For example, compounds with a
high affinity for protons, such as neuro-
transmitter amines, are selectively pro-
tonated by ammonia. Negative ion CI
often makes possible greatly enhanced
sensitivities for electronegative com-
pounds, such as those containing fluori-
nated, carboxyl, or nitro groups. Charge-
exchange CI can provide selectivity
based on ionization energy require-
ments; with benzene as the ionizing re-
agent gas one can selectively ionize aro-
matic compounds. The optimum concen-
tration of reagent gas relative to that of
the sample is 103; thus for solutions a
sensitivity increase of approximately this
amount can be obtained by using the
solvent as the ionizing reagent gas (28),
introducing the sample through an inter-
face used for liquid chromatography-
mass spectrometry (LC-MS) (9). Atmo-
spheric pressure ionization (API) pro-
vides high sensitivities for trace compo-
nents in air by a similar mechanism (29).
Selectivity in primary ion dissociation.
Decomposition of the separated primary
ions to produce the secondary mass
spectrum can occur spontaneously; how-
ever, such a metastable-ion MS-II spec-
Ms
lnterface He mm ESA-II
region
trum contains at most a few peaks, the
largest of which represent ions present in
only 10-2 to 10-4 times the abundance of
the precursor ion. To increase the num-
ber and absolute abundance of peaks in
the secondary mass spectrum, it is nec-
essary to add energy to the separated
primary ions; collisionally activated dis-
sociation (CAD) is the most convenient
and widely used method for doing this
(30, 31). Collision of the ions with gas
molecules (at a pressure of 10-4 torr)
_
converts part of the translational energy
of the ions into internal energy. The
subsequent unimolecular decomposition
of these excited ions is similar to that of
excited ions formed initially in the ion
source, and in general follows correla-
tion rules developed for normal El mass
spectra (7). For ions with energies of
many kilovolts, excitation results from a
grazing collision. For instruments such
as the quadrupole MS, in which the ions
have very low translational energies (5 to
20 eV), relatively large scattering angles
are necessary (billiard ball collisions).
Yost and Enke (21) found an ingenious
method to direct these scattered second-
ary ions into MS-Il by using a quadru-
pole with only a radio-frequency field,
the heart of the tandem quadrupole MS-
MS instrument.
The ions of the selected mlz value
exiting from MS-I may still represent
many components of the original sample,
so that selective excitation of the desired
ions in the collision process can also be
helpful. The average amount of energy
added by collision can be increased by
increasing the ion translational energy
Collector
slit l
j 340 mm
I
ueamw
monitor
'-slit 705 mm
210 mm
SCIENCE, VOL. 214
(32) or number of collisions (33), thus
increasing the relative abundance of sec-
ondary product ions from reactions with
higher energy requirements. Collisions
can convert negative ions to positive
ions, or singly charged ions to doubly
charged ions. Cooks and co-workers
showed that daughter ions requiring
similar energies for their formation can
be selected by collecting ions scattered
at a specific angle in multikilovolt colli-
sions (34). Bimolecular products of reac-
tions between the primary ion and the
collision gas can be observed when low-
energy collisions take place, but these
could lead to matrix effects in targeted
compound analysis. Use of tailored colli-
sion conditions to obtain additional se-
lectivity in MS-MS is a subject worth
further investigation.
Selectivity versus resolution of MS-I
and MS-II. When double-focusing mass
spectrometers are employed as MS-MS
instruments, an electrostatic analyzer is
used as MS-I (35, 36) or as MS-II (16).
The collisions produce secondary ions
with a range of translational energies, so
that the energy-analyzed spectra often
show incomplete resolution of neighbor-
ing peaks even at relatively low masses
[this problem has been solved with 40-
kV postcollision ion acceleration (37)].
The newer tandem quadrupole instru-
ments using low-energy collisions
achieve unit mass resolution in both MS-
I and MS-Il (21). The separation of iso-
baric multiplets such as N2 (m/z
28.0061), CO (27.9949), and C2H4
(28.0313) with a high-resolution mass
spectrometer is well known; the use of
such an instrument as MS-I (22) or MS-II
(38) or both (39) can give a dramatic
improvement in MS-MS specificity for
particular problems.
Sensitivity. The subpicogram sensitiv-
ity possible in normal mass spectrometry
is not substantially degraded in MS-MS;
the ion losses in CAD and MS-II separa-
tion are in part compensated by the
reduction in "chemical noise" achieved
by the second analyzer. The efficiency of
CAD for converting multikilovolt pri-
mary ions into secondary ions is 1
to > 10 percent (39) while efficiencies of
5 to 100 percent have been reported (40,
41) for the quadrupole-confined low-en-
ergy CAD method. The transmission ef-
ficiency of the added MS approaches 100
percent for magnetic and electrostatic
analyzers, and for the quadrupole ana-
lyzers with ions of lower mass. Although
there have been few studies to date of
the ultimate sensitivities possible with
MS-MS, subpicogram detection limits
have been reported (21, 41).
16 OCTOBER 1981
Speed. Even with the added collision
process and second analyzer, the time
between ion formation and collection in
MS-MS is < 10-3 second for most in-
struments. With a dedicated computer,
data acquisition times are on the order of
10-3 to >1 second per targeted compo-
nent, depending on the accuracy and
sensitivity desired as well as the efficien-
cy of the data system. Sample handling
can be the limiting factor in time require-
ments. With a direct introduction probe,
solid samples can often be introduced at
a rate of more than one per minute,
although the pump-out time for marginal-
ly volatile sample components can seri-
ously reduce this rate. Gaseous and liq-
uid (9, 28) samples can be introduced
directly, with sample streams analyzed
continuously or many batch samples an-
alyzed per minute with efficient sample
changers.
Instrumentation
Despite the newness of the MS-MS
field, certain types of instruments now
appear to be preferable for particular
analytical applications. For the determi-
nation of targeted compounds of low or
medium molecular weight in complex
mixtures the tandem quadrupole has
substantial advantages: unit resolution in
MS-I and MS-II, efficient CAD conver-
sion of primary to secondary ions, effi-
cient computer control for multiple-ion
monitoring and GC-MS-MS, and, with
mass production, possibly a future price
comparable to that of GC-MS.
For problems involving high molecular
weight compounds, the ion transmis-
sion of a magnetic instrument (8- to 10-
kV ions) is an order of magnitude higher
than that of a quadrupole at high mass
(m/z 500 to 1000 at unit resolution), an
effect that is multiplied for MS-MS.
However, this advantage is lost through
incomplete resolution if only an electro-
static analyzer is used as MS-I or MS-II,
without postcollision acceleration (37).
The tandem double-focusing instrument
(Fig. 2) (39) has the additional advantage
for high molecular weight samples of
high resolution in MS-I and MS-II, but it
is the most expensive MS-MS instru-
ment. New methods (10-12) make possi-
ble the ionization of compounds of mo-
lecular weight >3000; unit mass resolu-
tion in this range requires the use of a
double-focusing instrument. In addition,
instruments employing multikilovolt ion
collisions appear at present to be more
suitable for molecular structure elucida-
tion (22).
Magnetic and electrostatic analyzer
combinations. Double-focusing mass
spectrometers used for high resolution
and exact mass measurements contain a
tandem combination of magnetic (B) and
electrostatic (E) analyzers. These can
also be used as separate mass analyzers
for MS-MS, and most early studies were
performed with B-E ("reversed geome-
try") instruments of this kind (16, 30).
The magnetic analyzer is used as MS-I to
separate the desired primary ions, which
are then collisionally dissociated. The
kinetic energy of the resulting secondary
ions is directly proportional to their
mass, so that the electrostatic analyzer
can serve as MS-II. The most serious
deficiency of this system is the less-than-
unit resolution of the secondary ion
spectra; this can be improved by increas-
ing the ion translational energy before
(39) or after (37) collision. Linked scan-
ning (35, 36) of the two analyzers, either
B-E or E-B, makes possible unit mass
resolution in the secondary ion mass
spectra, but then the primary ion is poor-
ly resolved from its neighboring mass
peaks. The high kinetic energies of the
ions used lead to relatively high ion
transmission efficiencies for such analyz-
ers. Data collection efficiency is substan-
tially increased by the use of new lami-
nated field magnets that are capable of
scanning a mass decade in 0.2 second
(39, 42).
Triple-sector and tandem double-fo-
cusing instruments. The capabilities of
two-sector (B-E or E-B) instruments can
be increased by the addition of one or
two B or E sectors in tandem. Triple-
sector instruments of the configuration
E-B-(CAD)-E (22), B-E-(CAD)-B (38),
or B-E-(CAD)-quadrupole (42) make
possible high-resolution separation of
the primary ions in MS-I; B-(CAD)-E-B
gives unit resolution in MS-I and high
resolution in MS-Il (38, 42). Use of a
fourth sector, as in an E-B-(CAD)-E-B
configuration (Fig. 2) (39), gives high
resolution in both MS-I and MS-II. A
double-focusing MS focuses for energy
as well as divergence; thus the MS-Il of
this instrument brings secondary ions of
different energies but the same mass
together at the same focal point, which
greatly narrows the resulting peaks and
increases their height and signal-to-noise
ratio. Commercially available multiple-
sector instruments cost $500,000 to
$700,000 (42, 43).
Tandem quadrupole instruments.
These mass analyzers (Fig. 1) provide
unit mass resolution in both MS-I and
MS-II, and the radio frequency-only
quadrupole collision region provides a
283
high-efficiency interface between them. tween the accelerating region and the
The unique characteristics of quadru- magnetic field of a single-focusing mass
poles which led to their widespread use spectrometer, but the resolution is poor
in GC-MS systems are similarly advanta- and interferences can be serious. Louter
geous in MS-MS. Quadrupoles are com- et al. (37) built a special instrument with
pact, relatively easy to maintain, and can greatly improved resolution for MS-II
be efficiently controlled by an on-line spectra; a channel plate array detector in
computer or microprocessor. Such in- MS-II allows continuous measurement
struments are available commercially and display of these secondary spectra,
from several sources (44-46), and per- which is valuable for monitoring fast
haps the rapidly growing market will experiments such as pulsed laser ioniza-
substantially reduce their price; at pre- tion. Glish et al. (47) constructed a mag-
sent, the instrument with a data system netic-quadrupole system with consider-
costs $350,000 to $450,000. able versatility, which is useful for re-
Other MS-MS instruments. Mass search-for instance, in studies of the
spectral information can actually be ob- effect of energy on the low-energy colli-
tained from metastable or collisionally sion process. McIver (48) and Freiser
activated decompositions that occur be- (49) have proposed MS-MS instruments
Table 1. Examples of MS-MS applications.
System studied Principal investigator
Chemotaxonomy of cacti R. G. Cooks, Purdue University
Drugs in raw urine D. F. Hunt, University of Virginia
Polynuclear aromatics C. G. Enke, Michigan State University
Ion photodissociation J. D. Morrison and D. C. McGilvery, LaTrobe
University, Melbourne
Tetrachlorodibenzodioxin M. L. Gross, University of Nebraska
DNA pyrolysis K. Levsen, Bonn University
Marine sterols A. Maquestiau, Mons University, Belgium
Polymer sequencing P. J. Derrick, LaTrobe University, Melbourne
Aromatic amines K. R. Jennings, University of Warwick, England
Steroid mixtures C. Djerassi, Stanford University
Saxotoxin A. L. Burlingame, University of California, Berkeley
Phosphate ester chirality J. R. Knowles, Harvard University
Ion structures H. Schwarz, Technical University of Berlin,
West Germany
Nitroaromatics J. H. Beynon, University College of Swansea, Wales
Stereoisomers J. Marsel, University of Ljubljana, Yugoslavia
Natural product pyrolyses H. L. C. Meuzelaar, University of Utah
Pyrolysis of bacteria A. J. H. Boerboom, FOM-Instituut voor Atoom- en
Molecuulfysica, Amsterdam
Short-lived radicals C. N. McEwen, DuPont
Alkaloids in plants W. F. Haddon, U.S. Department of Agriculture
Insect pheromones C. G. MacDonald, M. T. Lacey, CSIRO, Canberra
Isobaric mixtures D. Stahl, Federal Technical Institute, Lausanne
Protein digests H. R. Morris, Imperial College, London
Penicillins J. Occolowicz, Eli Lilly Co.
Trace contaminants in air R. A. Yost, University of Florida
J. R. Hass, National Institute of Environmental
Health Sciences
Benzopyrenes G. A. McClusky, Frederick Cancer Research Center
Petrochemicals R. W. Kondrat, Shell Development Co.
Metriprananol in serum M. Senn, Boehringer-Mannheim
Substituted aromatics D. H. Russell, Oak Ridge National Laboratory
Metabolic profiling M. Anbar, State University of New York, Buffalo
Polychlorinated biphenyls R. K. Boyd, University of Guelph
Drug metabolites A. Tatematsu, Meijo University, Japan
Detection of explosives J. Yinon, Weizmann Institute, Israel
Enkephalins R. M. Caprioli, University of Texas Medical School
Polymer additives I. Sakai, Toray Industries, Japan
Coal liquids B. W. Wilson, Battelle Northwest Laboratories
Prostaglandins W. K. Duholke, Upjohn Co.
Steranes in crude oil E. J. Gallegos, Chevron Research
Peptide mixtures R. R. Ragakov, Tashkent State University, U.S.S.R.
Testosterone in tissue S. J. Gaskell, Welsh National School of Medicine
Diesel exhaust D. Schuetzle, Ford Motor Co.
Odors in air V. J. Caldecourt, Dow Chemical Co.
Concealed drugs, fruits J. B. French, Sciex, Inc., Toronto
Parathion in lettuce J. R. B. Slayback, Finnigan Corp.
Ion plasmas M. W. Siegel, Extranuclear Laboratories, Inc.
Sequences of mixed peptides F. W. McLafferty, Cornell University
284
based on the ion cyclotron resonance
(ICR) principle. On the basis of experi-
ence with our earlier instrument (50), we
have designed a tandem time of flight
mass spectrometer with a two-dimen-
sional array detector. A variety of tan-
dem mass spectrometers, including mag-
netic-ICR and tandem double-focusing
instruments, have been constructed for
other purposes, such as studies of ion-
molecule reactions and trace analysis for
nuclidic species (22-24).
Applications
There has been a great increase in the
last 2 years in the number of laboratories
utilizing MS-MS and in the number and
types of problems to which it has been
applied. A number of examples are
shown in Table 1; many more are given
in recent reviews (17-25).
One of the first laboratories active in
MS-MS research was that of Cooks and
co-workers; their work attracted wide-
spread attention to the potential of the
technique. For example, they performed
direct quantitative analyses of urine for
picogram amounts of homovanillic acid
and testosterone; untreated samples
were dried on the MS insertion probe
and the components vaporized directly
in the ion source (19). Similarly, they
analyzed stems, leaves, and berries of
coca plant for cocaine and cinnamoylco-
caine by direct ion source introduction of
1-mm3 samples (20). Recently, they
showed that direct MS-MS analysis of
samples of low volatility, such as saccha-
rides and quaternary ammonium salts, is
possible with laser desorption ionization
(51).
Hunt and his colleagues have per-
formed a wide variety of analyses with a
tandem quadrupole MS-MS instrument
using negative (52) as well as positive
ions. The compounds analyzed included
90 carboxylic acids in urine, peptide mix-
tures with component sequencing, or-
ganosulfur compounds in crude oil and
coal liquids, and pollutants present in
sewage sludge at a concentration of 100
parts per billion (the total anWlysis time
was 15 minutes per sample) (52, 53).
Enke and co-workers (54) obtained re-
sults that show promise for routine anal-
ysis of the acid fraction components of
priority pollutants in complex mixtures.
Meuzelaar et al. (55) have used MS-
MS in the classification of complex ma-
terials such as body tissues, bacteria,
and coal by analyzing the distribution of
products formed on Curie-point pyroly-
sis of submicrogram samples. Haddon
SCIENCE, VOL. 214
and Molyneux (56) analyzed samples of
Senecio and and similar plants for pyrro-
lizidine alkaloids, which are toxic to live-
stock, and found the method 100 times
more sensitive and 30 times faster than
the routine LC method used previously.
Slayback and Story (25) obtained linear
calibration curves down to 10-11 g for
the pesticide parathion in ivy and lettuce
leaves; although the negative ion CI
spectrum of the crude uncontaminated
leaf showed a peak at virtually every
mass, there was no interference for the
characteristic secondary ions at m/z 154
and 169 formed from the primary ion at
m/z 291. The API technique is highly
sensitive and convenient for direct sam-
pling. Caldecourt et al. (57) used API to
measure odor components in air at con-
centrations of parts per billion and deter-
mined the structures of unknown com-
pounds; some matrix effects on quantita-
tion were also noted. Davidson et al. (41)
described a variety of applications of
API, including analyses of vapors from
drugs, explosives, and fruits in air ex-
pressed from luggage; nitrosamines in air
above frying bacon or in beer; polynucle-
ar aromatics in oil; aflatoxin B1 in peanut
butter; sulfamethazine in pork; and drug
metabolites in raw urine. The quantita-
tive response for tetrachlorodibenzo-
dioxin in fish extracts was found to be
linear to less than 10-12 g.
High molecular weight samples. Lins-
cheid et al. (58) obtained structural infor-
mation for polysaccharide mixtures by
field desorption ionization and MS-MS;
mycobacterial methylmannose polysac-
charides were characterized up to m/z
2506. Barbalas et al. (59) applied MS-MS
to polybromobiphenyls and to the cardi-
ac glycosides digoxin and digitoxin. Fast
atom bombardment ionization and MS-
MS were used by Morris et al. (60) to
analyze mixtures of polar and nonpolar
compounds of high molecular weight,
such as carbohydrates and peptides. A
new high-field (2.3-tesla) magnet of radi-
us 30 cm allowed them to use 8-kV ions
up to m/z 3000. At Cornell we are install-
ing a 2.3-tesla magnet of radius 60 cm on
the tandem double-focusing MS-MS in-
strument (39), which should allow mass
analysis of such ions up to m/z 12,000.
Use of MS-MS with separation sys-
tems. Tandem mass spectrometers can
also act as a very specific detector for
GC and other separation systems, in-
cluding the crude fractionation obtained
by heating the sample in the MS ion
source. For such use rapid scanning and
data acquisition are valuable, so that the
computer-controlled tandem quadrupole
instruments are often particularly appro-
16 OCTOBER 1981
100 ,I4
I vfoa
50 - 30
20
o0
Fig. 3. Secondary collision mass spectra produced from fragment ions in the electron ionization
mass spectrum of 5-methyl-3-hexanone.
priate. In these applications MS-MS can
be used in three general modes, to sup-
ply (i) a secondary ion spectrum that is
characteristic of a specific primary ion
(also used for targeted compound analy-
sis); (ii) primary ions that yield a com-
mon secondary ion (or ions) which is
characteristic of a structural class; and
(iii) primary ions that yield secondary
ions through loss of a neutral moiety (or
moieties) which is characteristic of a
specific structural class.
The first mode is illustrated by the
identification of penicillins in mixtures
such as fermentation broths. When ion-
ized by electrons, these compounds form
abundant C7H12NO2S+ fragment ions
from their common tetrahydrothiazole
ring; after collision, these primary ions
yield a secondary mass spectrum charac-
teristic of penicillins (61). This allows
penicillins to be distinguished from other
compounds, such as dihydrocephalo-
sporins and tetramethylthiazolidine car-
boxylates, which also give large
C7H12NO2S' peaks in their El spectra.
As an example of the second mode,
Gallegos (62) reported that terpane and
sterane molecular ions produce abun-
dant secondary peaks at m/z 191 and 217,
respectively, on metastable decomposi-
tion. A sample from Green River shale
vaporized from the direct probe pro-
duced a homologous series (C18 to C35)
of molecular ion peaks yielding second-
ary peaks at m/z 191 and another homol-
ogous series yielding peaks at m/z 217.
By assuming that relative secondary ion
yields were the same for all homologs,
Gallegos obtained quantitative values in
Secondary
spectra
agreement with those derived by a much
more time-consuming analysis. Two ex-
amples of the third mode may be given.
Zakett et al. (63) used the loss of mass 44
(CO2) from primary ions (M - 1)- to
indicate that those ions originated from
carboxylic acids, and Shushan et al. (64)
found that M+ ions of polychlorobiphen-
yls and other highly chlorinated com-
pounds are indicated by their character-
istic secondary losses of Cl and C12.
Molecular structure determination.
Secondary CAD spectra of multikilovolt
ions can be characteristic of the struc-
ture of the ion, independent of its mode
of formation (30, 31, 65) [although in low-
energy CAD spectra variations with
mode of formation are observed (54)].
The molecular structure of a pure com-
pound can thus be determined by identi-
fying fragment ion structures in its nor-
mal EI mass spectrum. Figure 3 shows
MS-MS information derived from 5-
methyl-3-hexanone. The primary high-
resolution mass spectrum shows two
peaks of nominal mass 57, C3H50 and
C4H9+. The secondary CAD spectrum of
the former agrees well with a reference
spectrum of the C2H5CO' ion; if a refer-
ence spectrum had not been available it
might have been identified from the char-
acteristic loss of CO to form the abun-
dant C2H5' peak. The CAD spectra of
the C4H9+ isomers are the same within
experimental error. However, the pres-
ence of the m/z 72 peak, as well as its
CAD spectrum, shows that the 3-posi-
tion is unsubstituted.
In collaboration with D. A. Cooper of
the U.S. Drug Enforcement Administra-
285

Fig. 4. Cathode-ray tube display of the results of computer interpretation of the data of Fig. 3.
The fragment ion structures at the bottom of the screen were identified by probability-based
computer matching of the observed CAD spectra against the reference file. The data at the
lower right are possible combinations of these substructures that fit the proposed molecular
formula C7HI40. One possible molecular structure consistent with these data, that of 3-
heptanone, was assembled by the interpreter with a light pen accessory.
tion, M. T. Cheng and G. H. Kruppa of combinations of 3-OH group configura-
Cornell used MS-MS to examine the tions and A-B ring geometry: 5a,3a;
street drug "china white," which ap- 5a,33; 5P,3x; and 5P3,3p3. These CAD
peared in the Los Angeles area in 1980. reference spectra are useful for charac-
Cooper and his associates had assigned terizing the stereochemistry as well as
the structure shown below (1) on the the primary structure of unknown com-
0 pounds; the CAD spectrum from the m/z
11 ,C2H5 234 peak of 5a-cholestane-3a-ol is the
CH2-CH(CH3)-N } N
same within experimental error as the
3a,5a reference spectrum.
It would be advantageous to be able to
identify substructures automatically by
basis of MS, NMR spectroscopy, infra- comparison of fragment ion CAD spectra
red spectroscopy, and synthesis. Its with a reference collection. To this end,
mass spectrum shows prominent peaks we have collected a data base of 700
at several masses, including 57, 91, and CAD spectra (average mass, 110) (66).
146. Comparison of the peaks with refer- To search this file, I. K. Mun, M. P.
ence CAD spectra indicate the substruc- Barbalas, and W. Staedeli of Cornell
tures C2H5CO', C6H5CH2', and developed an algorithm modeled after
C6H5NH+=C(CH3)CHCH2 (by rear- the probability-based matching system
rangement), which would be sufficient to for EI mass spectra (67). The computer
determine the structure. can also serve as an aid to construct
Such secondary spectra can also pro- possible molecules consistent with the
vide stereochemical information. M. T. identified substructures, molecular ion
Cheng, M. P. Barbalas, and R. F. Pegues composition, and other information. Fig-
at Cornell studied a variety of 3- ure 4 shows a cathode-ray tube display
hydroxysteroids that produce a charac- of results of such an interpretation of the
teristic mass spectral peak at mlz 234 data in Fig. 3; the operator assembled
corresponding to an ion containing the the retrieved substructures by using light
A, B, and C rings as shown below (2). pen manipulation and showed that the
R structure of 2-heptanone is consistent
with the data.
Future Prospects
XS 234
Ho
H
Two general areas of MS-MS growth
2 N. Tyler, J. Chem. Soc. Chem. Commun. 1981,
appear to be particularly promising. Use
Using pregnane-3,20-diols, they found of MS-MS for routine analysis for lower
four different secondary CAD spectra of molecular weight compounds in complex
these m/z 234 ions for the four possible mixtures with a minimum of sample
286
work-up should continue to increase. I
believe that this demand will bring the
price of computerized MS-MS instru-
ments down to $50,000; the present price
appears to be the major deterrent to MS-
MS growth rivaling that of GC-MS and
LC-MS in recent years. The versatility
of such an instrument should make it
applicable to a wide variety of problems
in industrial, clinical, regulatory, and re-
search laboratories, where it could per-
form 1000 or more quantitative analyses
per day. This is already possible in nor-
mal mass spectrometry with direct liquid
introduction systems (9) and commercial
sample changers; instrument reliability
will be a critical factor in achieving such
a goal.
A second area in which MS-MS ap-
pears to have a high potential is the
analysis and characterization of macro-
molecules, including biomedical sam-
ples, plant materials, crude oil, and com-
mercial polymers. A key problem is that
of obtaining mass spectral information
from nonvolatile compounds (10-12, 68).
When further progress is made in this
direction it may be possible to introduce
complex biological material into the ion
source of a tandem double-focusing MS-
MS (39) and use CI to selectively pro-
duce (M + H)+ ions, corresponding to
polypeptides and other species with a
high proton affinity, for unit mass sepa-
ration in MS-I; this should be possible to
m/z 50,000 for 2-kV ions. The secondary
CAD spectrum of such a polypeptide ion
could provide complete information
about its amino acid sequence; this can
be done for peptides of molecular weight
2000 (60). It may be possible to do this
with picomole quantities of material, al-
though the sensitivity is highly depen-
dent on a number of to-be-determined
factors, such as ionization efficiency and
computer-controlled data acquisition.
References and Notes
I. R. S. Yalow, Science 200, 1236 (1978).
2. G. H. Fricke, Anal. Chem. 52, 259R (1980).
3. C. R. Martin and H. Freiser, ibid., p. 562.
4. 0. H. Lowry, J. V. Passonneau, D. W. Schulz,
M. K. Rock, J. Biol. Chem. 236, 2746 (1961).
5. E. L. Wehry, Anal. Chem. 52, 75R (1980).
6. B. L. Karger, L. R. Snyder, C. Horvath, Intro-
duction to Separation Sciences (Wiley, New
York, 1973).
7. F. W. McLafferty, Interpretation of Mass Spec-
tra (University Science Books, Mill Valley, Cal-
if., ed. 3, 1980).
8. C. J. W. Brooks and C. G. Edmonds, in Practi-
cal Mass Spectrometry, B. S. Middleditch, Ed.
(Plenum, New York, 1979), pp. 57-126.
9. B. G. Dawkins and F. W. McLafferty, in GLC
and HPLC Determination of Therapeutic
Agents, K. Tsuji and W. Morozowich, Eds.
(Dekker, New York, 1978), vol. 1, pp. 259-275.
10. F. W. McLafferty and E. R. Lory, J. Chroma-
togr. 203, 109 (1981).
11. M. Barber, R. S. Bordoli, R. D. Sedgwick, A.
325 (1981).
12. C. J. McNeal and R. D. Macfarlane, J. Am.
Chem. Soc. 103, 1609 (1981).
13. F. W. McLafferty and F. M. Bockhoff, Anal.
Chem. 50, 69 (1978).
SCIENCE, VOL. 214
 14. W. F. Haddon, in High Performance Mass
Spectrometry, M. L. Gross, Ed. (American
Chemical Society, Washington, D.C., 1978), pp.
97-119.
15. F. W. McLafferty and T. A. Bryce, J. Chem.
Soc. Chem. Commun. 1967, 1215 (1967).
16. R. G. Cooks, J. H. Beynon, R. M. Caprioli, G.
R. Lester, Metastable Ions (Elsevier, Amster-
dam, 1973).
17. H.-K. Wipf, P. Irving, M. McCamish, R. Ven-
kataraghavan, F. W. McLafferty, J. Am. Chem.
Soc. 95, 3369 (1973).
18. D. H. Smith, C. Djerassi, K. H. Maurer, U.
Rapp, ibid. 96, 3482 (1974).
19. T. L. Kruger, J. F. Litton, R. W. Kondrat, R. G.
Cooks, Anal. Chem. 48, 2113 (1976).
20. R. W. Kondrat and R. G. Cooks, ibid. 50, 81A
(1978).
21. R. A. Yost and C. G. Enke, ibid. 51, 1251A
(1979).
22. F. W. McLafferty, Accounts Chem. Res. 13, 33
(1980).
23. T. H. Maugh II, Science 209, 675 (1980).
24. P. F. Bente, III, and F. W. McLafferty, Pract.
Spectrosc. 3, 253 (1980).
25. J. R. B. Slayback and M. S. Story, Ind. Res.
Dev. 1981, 129 (February 1981).
26. K. Levsen, Adv. Mass Spectrom. 8, 897 (1980).
27. D. F. Hunt, G. C. Stafford, Jr., F. W. Crow, J.
W. Russell, Anal. Chem. 48, 2098 (1976).
28. P. Price, D. P. Martinsen, R. A. Upham, H. S.
Swofford, Jr., S. E. Buttrill, Jr., ibid. 47, 190
(1975).
29. E. L. Horning, M. G. Horning, P. I. Carroll, I.
Dzidic, R. N. Stillwell, ibid. 45, 936 (1973).
30. F. W. McLafferty, P. F. Bente, III, R. Kornfeld,
S.-C. Tsai, I. Howe, J. Am. Chem. Soc. 95,2120
(1973).
31. F. W. McLafferty, Philos. Trans. R. Soc. Lon-
don Ser. A 293, 93 (1979).
32. M. S. Kim and F. W. McLafferty, J. Am. Chem.
Soc. 100, 3279 (1978).
33. P. J. Todd and F. W. McLafferty, Int. J. Mass
Spectrom. Ion Phys. 38, 371 (1981).
34. A. R. Hubik, P. H. Hemberger, J. A. Laramee,
R. G. Cooks, ibid. 102, 3997 (1980).
Bioselective Membrane
Electrode Probes
One of the most rapidly expanding
research areas relating to analytical mea-
surements is the development of poten-
tiometric membrane electrodes with se-
lectivity for ions, dissolved gases, and
biological materials. Activity in this field
is so intense that new publications are
appearing at a rate approaching 500 per
year (1). Entirely aside from the appeal-
ing practical possibilities for such poten-
tiometric membrane electrodes, it ap-
pears that new research directions have
been directly stimulated by the timely
infusion of concepts from physics (2)
and biology (3). Some of the conse-
quences of the latter are examined in this
article.
SCIENCE, VOL. 214, 16 OCTOBER 1981
35. A. F. Weston, K. R. Jennings, S. Evans, R. M.
Elliott, Int. J. Mass Spectrom. Ion Phys. 20, 317
(1976).
36. B. Shushan and R. K. Boyd, Anal. Chem. 53,
421 (1981).
37. G. J. Louter, A. J. H. Boerboom, P. F. M.
Stalmeier, H. H. Tuithof, J. Kistemaker, Int. J.
Mass Spectrom. Ion Phys 33, 335 (1980).
38. F. W. McLafferty and P. J. Todd, Org. Mass
Spectrom. 15, 272 (1980).
39. F. W. McLafferty, P. J. Todd, D. C. McGilvery,
M. A. Baldwin, J. Am. Chem. Soc. 102, 3360
(1980).
40. R. A. Yost, paper presented at the Summer
Symposium, Analytical Division, American
Chemical Society, Pittsburgh, Pa., July 1981.
41. W. R. Davidson, J. Fulford, N. M. Reid, T.
Sakuma, B. Shushan, B. A. Thomson, paper
presented at the American Society of Mass
Spectrometry (ASMS) meeting, Minneapolis,
Minn., May 1981.
42. VG-Micromass, Altrincham, Cheshire, En-
gland.
43. Kratos, Ltd., Urmston, Manchester, England.
44. Sciex, Inc., Thormhill, Ontario, Canada.
45. Finnigan Corp., Sunnyvale, Calif.
46. Extranuclear, Inc., Pittsburgh, Pa.
47. G. L. Glish, D. Zackett, P. H. Hemberger, R. G.
Cooks, paper presented at the ASMS meeting,
New York, May 1980.
48. R. T. McIver, personal communication.
49. B. S. Freiser, personal communication.
50. W. F. Haddon and F. W. McLafferty, Anal.
Chem. 41, 31 (1969).
51. D. Zackett, A. E. Schoen, R. G. Cooks, P. H.
Hemberger, J. Am. Chem. Soc. 103, 1295
(1981).
52. D. F. Hunt, J. Shabanowitz, A. B. Giordani,
Anal. Chem. 52, 386 (1980).
53. D. F. Hunt, J. Shabanowitz, A. B. Giordani, T.
M. Harvey, paper presented at the ASMS meet-
ing, Minneapolis, May 1981.
54. J. A. Chakel, C. A. Myerholtz, C. G. Enke,
paper presented at the ASMS meeting, Minne-
apolis, May 1981.
55. H. L. C. Meuzelaar, W. H. McClennan, G. S.
Garry A. Rechnitz
Potentiometric membrane electrodes
have their conceptual origin in the ubiq-
uitous pH glass membrane electrode
which, in its present form, is still the
most sensitive and selective of all such
electrodes. During the past 20 years, a
wide variety of conventional potentio-
metric ion-selective membrane elec-
trodes based on glasses (4), crystals (5),
and various liquid membranes (6-8) has
been developed and commercialized.
Several recent reviews (1) and mono-
graphs (9, 10) give comprehensive ac-
counts of this work.
Throughout the period of ion-selective
electrode development, efforts were
made to extend the measurement capa-
0036-8075/81/1016-0287$01.00/0 Copyright © 1981 AAAS
Metcalf, and G. R. Hill, paper presented at the
ASMS meeting, Minneapolis, May 1981.
56. W. F. Haddon and R. Molyneux, paper present-
ed at the MS-MS Conference, Asilomar, Calif.,
September 1980.
57. V. J. Caldecourt, D. Zackett, J. C. Tou, paper
presented at the ASMS meeting, Minneapolis,
May 1981.
58. M. Linscheid, J. D'Angona, A. L. Burlingame,
A. Dell, C. A. Ballou, Proc. Natl. Acad. Sci.
U.S.A. 78, 1471 (1981).
59. M. P. Barbalas, M. T. Cheng, C. Wesdemiotis,
I. J. Amster, C. J. Sack, F. W. McLafferty,
paper presented at the ASMS meeting, Minne-
apolis, May 1981.
60. H. R. Morris, M. Panico, M. Judkins, A. Dell,
R. McDowell, paper presented at the ASMS
meeting, Minneapolis, May 1981.
61. J. L. Occolowitz, M. P. Barbalas, F. W. McLaf-
ferty, paper presented at the ASMS meeting,
New York, May 1980.
62. E. J. Gallegos, Anal. Chem. 48, 1348 (1976).
63. D. Zakett, A. E. Schoen, R. W. Kondrat, R. G.
Cooks, J. Am. Chem. Soc. 101, 6783 (1979).
64. B. Shushan, N. J. Bunce, R. K. Boyd, C. T.
Corke, Biomed. Mass Spectrom. 8, 225 (1981).
65. F. W. McLafferty, A. Hirota, M. P. Barbalas,
R. F. Pegues, Int. J. Mass Spectrom. Ion Phys.
35, 299 (1980).
66. F. W. McLafferty, A. Hirota, M. P. Barbalas,
Org. Mass Spectrom. 15, 327 (1980).
67. G. M. Pesyna, R. Venkataraghavan, H. E.
Dayringer, F. W. McLafferty, Anal. Chem. 48,
1362 (1976).
68. P. J. Todd and G. L. Glish at Oak Ridge
National Laboratory have used secondary ion
bombardment for MS-MS analysis of samples of
low volatility (personal communication).
69. Discussions with I. K. Mun and M. P. Barbalas
were particularly helpful in preparing this arti-
cle. The generous financial support of the MS-
MS research program at Cornell by the National
Institutes of Health (grant GM16609) and the
Army Research Office, Durham (grant
DAAG29-79-C-0046), is gratefully acknowl-
edged.
bilities of membrane electrodes to bio-
logical materials through the use of en-
zyme catalysis to convert substrates to
species that could be sensed by ion-
selective membrane electrodes. The re-
sulting "enzyme electrodes" represent
an increasingly practical, but now largely
conventional, means of effecting biose-
lective measurements with membrane
electrodes (11).
A special impetus was given to re-
search on bioselective membrane elec-
trodes in the early 1970's when stable
and reliable potentiometric sensors for
ammonia, carbon dioxide, hydrogen sul-
fide, and other dissolved gases became
commercially available on a routine ba-
sis. Such electrodes combine the tech-
nology of ion-selective membrane elec-
trodes with that of microporous synthet-
ic membranes (12). The ammonia gas-
sensing membrane electrode, for exam-
ple, is a potentiometric sensor in which a
hydrophobic gas-permeable membrane
is superimposed on a flat pH-type glass
membrane electrode in contact with a
thin layer of ammonium chloride electro-
lyte solution. This arrangement gives ex-
ceptional selectivity for the measure-
The author is Unidel Professor of Chemistry,
University of Delaware, Newark 19711.
287